Beamforming technology, which effectively performs wireless signal transmission by an antenna array, is often applied in wireless communication to achieve the purposes such as enhancement of signal quality, reduction of interference, improvement of transmission efficiency and so on. In a multiple-input multiple-output (MIMO) system with massive antennas (for example, 16, 64 or more antennas), the antenna array may be used to effectively compensate path fading due to distance and to enhance performance of the MIMO system. Such antenna array is conventionally implemented by an all-digital architecture. That is, each antenna in the antenna array requires a digital signal processing circuit corresponding thereto. For example, each antenna generally corresponds to one transmitter (TX) chain, and each TX chain at least includes a variety of high-end circuit elements such as a digital-to-analog converter (DAC), a filter, a power amplifier (PA) and so on. However, these signal processing circuits are of high cost and have higher power consumption.
To reduce the cost of establishing the beamforming technology, in recent hybrid beamforming systems, the signal processing circuits have been transformed from the all-digital form into a mixed structure of analog and digital circuits, in which some circuits (for example, TX chains or other elements) have been reduced in number, thereby reducing the establishment cost. Nevertheless, in the analog circuit, it is difficult to fine-tune a phase shifter and beamforming gain. As a result, in actual operation, the hybrid beamforming system is not as easy to use as a beamforming system implemented by the all-digital architecture. From another point of view, the analog circuit limitations in hybrid beamforming form a very complex non-convex optimization problem to obtain the optimal solution, which is disadvantageous for practical applications.